บทเรียนนี้จะครอบคลุม:
- การสำรวจ Mistral Models ที่แตกต่างกัน
- การเข้าใจกรณีการใช้งานและสถานการณ์สำหรับแต่ละโมเดล
- ตัวอย่างโค้ดที่แสดงคุณสมบัติเฉพาะของแต่ละโมเดล
ในบทเรียนนี้ เราจะสำรวจ 3 โมเดล Mistral ที่แตกต่างกัน:
Mistral Large, Mistral Small และ Mistral Nemo
โมเดลเหล่านี้ทั้งหมดสามารถใช้งานได้ฟรีบน Github Model marketplace โค้ดในโน้ตบุ๊กนี้จะใช้โมเดลเหล่านี้ในการรันโค้ด รายละเอียดเพิ่มเติมเกี่ยวกับการใช้ Github Models เพื่อ สร้างต้นแบบด้วย AI models
Mistral Large 2 เป็นโมเดลเรือธงจาก Mistral ที่ออกแบบมาเพื่อการใช้งานในองค์กร
โมเดลนี้เป็นการอัปเกรดจาก Mistral Large รุ่นแรก โดยมี
- ขนาดหน้าต่างบริบทที่ใหญ่ขึ้น - 128k เทียบกับ 32k
- ประสิทธิภาพที่ดีขึ้นในงานคณิตศาสตร์และการเขียนโค้ด - ความแม่นยำเฉลี่ย 76.9% เทียบกับ 60.4%
- ประสิทธิภาพหลายภาษาเพิ่มขึ้น - รองรับภาษาอังกฤษ ฝรั่งเศส เยอรมัน สเปน อิตาลี โปรตุเกส ดัตช์ รัสเซีย จีน ญี่ปุ่น เกาหลี อาหรับ และฮินดี
ด้วยคุณสมบัติเหล่านี้ Mistral Large โดดเด่นในด้าน
- Retrieval Augmented Generation (RAG) - เนื่องจากมีหน้าต่างบริบทที่ใหญ่ขึ้น
- Function Calling - โมเดลนี้รองรับการเรียกฟังก์ชันโดยตรง ทำให้สามารถเชื่อมต่อกับเครื่องมือและ API ภายนอกได้ การเรียกฟังก์ชันสามารถทำได้ทั้งแบบขนานหรือเรียงลำดับทีละคำสั่ง
- Code Generation - โมเดลนี้โดดเด่นในการสร้างโค้ด Python, Java, TypeScript และ C++
ในตัวอย่างนี้ เราใช้ Mistral Large 2 เพื่อรันรูปแบบ RAG บนเอกสารข้อความ คำถามถูกเขียนเป็นภาษาเกาหลีและถามเกี่ยวกับกิจกรรมของผู้เขียนก่อนเข้ามหาวิทยาลัย
โมเดลใช้ Cohere Embeddings Model เพื่อสร้าง embeddings ของเอกสารและคำถาม สำหรับตัวอย่างนี้ใช้แพ็กเกจ faiss ของ Python เป็นที่เก็บเวกเตอร์
คำสั่งที่ส่งไปยังโมเดล Mistral รวมทั้งคำถามและส่วนที่ถูกดึงมา (retrieved chunks) ที่มีความคล้ายคลึงกับคำถาม โมเดลจึงให้คำตอบในรูปแบบภาษาธรรมชาติ
pip install faiss-cpuimport requests
import numpy as np
import faiss
import os
from azure.ai.inference import ChatCompletionsClient
from azure.ai.inference.models import SystemMessage, UserMessage
from azure.core.credentials import AzureKeyCredential
from azure.ai.inference import EmbeddingsClient
endpoint = "https://models.inference.ai.azure.com"
model_name = "Mistral-large"
token = os.environ["GITHUB_TOKEN"]
client = ChatCompletionsClient(
endpoint=endpoint,
credential=AzureKeyCredential(token),
)
response = requests.get('https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/paul_graham/paul_graham_essay.txt')
text = response.text
chunk_size = 2048
chunks = [text[i:i + chunk_size] for i in range(0, len(text), chunk_size)]
len(chunks)
embed_model_name = "cohere-embed-v3-multilingual"
embed_client = EmbeddingsClient(
endpoint=endpoint,
credential=AzureKeyCredential(token)
)
embed_response = embed_client.embed(
input=chunks,
model=embed_model_name
)
text_embeddings = []
for item in embed_response.data:
length = len(item.embedding)
text_embeddings.append(item.embedding)
text_embeddings = np.array(text_embeddings)
d = text_embeddings.shape[1]
index = faiss.IndexFlatL2(d)
index.add(text_embeddings)
question = "저자가 대학에 오기 전에 주로 했던 두 가지 일은 무엇이었나요??"
question_embedding = embed_client.embed(
input=[question],
model=embed_model_name
)
question_embeddings = np.array(question_embedding.data[0].embedding)
D, I = index.search(question_embeddings.reshape(1, -1), k=2) # distance, index
retrieved_chunks = [chunks[i] for i in I.tolist()[0]]
prompt = f"""
Context information is below.
---------------------
{retrieved_chunks}
---------------------
Given the context information and not prior knowledge, answer the query.
Query: {question}
Answer:
"""
chat_response = client.complete(
messages=[
SystemMessage(content="You are a helpful assistant."),
UserMessage(content=prompt),
],
temperature=1.0,
top_p=1.0,
max_tokens=1000,
model=model_name
)
print(chat_response.choices[0].message.content)Mistral Small เป็นอีกหนึ่งโมเดลในตระกูล Mistral ที่อยู่ในกลุ่ม premier/enterprise ตามชื่อ โมเดลนี้เป็น Small Language Model (SLM) ข้อดีของการใช้ Mistral Small คือ
- ประหยัดค่าใช้จ่ายเมื่อเทียบกับ Mistral LLMs เช่น Mistral Large และ NeMo - ลดราคาลง 80%
- ความหน่วงต่ำ - ตอบสนองได้เร็วกว่า LLMs ของ Mistral
- ยืดหยุ่น - สามารถติดตั้งในสภาพแวดล้อมต่าง ๆ ได้ง่ายขึ้นโดยมีข้อจำกัดด้านทรัพยากรน้อยกว่า
Mistral Small เหมาะสำหรับ:
- งานที่ใช้ข้อความ เช่น การสรุปความ วิเคราะห์ความรู้สึก และการแปลภาษา
- แอปพลิเคชันที่มีการร้องขอบ่อย ๆ เนื่องจากประหยัดค่าใช้จ่าย
- งานโค้ดที่ต้องการความหน่วงต่ำ เช่น การตรวจสอบและแนะนำโค้ด
เพื่อแสดงความแตกต่างของความหน่วงระหว่าง Mistral Small และ Large ให้รันเซลล์ด้านล่าง
คุณจะเห็นความแตกต่างของเวลาตอบสนองประมาณ 3-5 วินาที และสังเกตความยาวและสไตล์ของคำตอบจากคำสั่งเดียวกัน
import os
endpoint = "https://models.inference.ai.azure.com"
model_name = "Mistral-small"
token = os.environ["GITHUB_TOKEN"]
client = ChatCompletionsClient(
endpoint=endpoint,
credential=AzureKeyCredential(token),
)
response = client.complete(
messages=[
SystemMessage(content="You are a helpful coding assistant."),
UserMessage(content="Can you write a Python function to the fizz buzz test?"),
],
temperature=1.0,
top_p=1.0,
max_tokens=1000,
model=model_name
)
print(response.choices[0].message.content)import os
from azure.ai.inference import ChatCompletionsClient
from azure.ai.inference.models import SystemMessage, UserMessage
from azure.core.credentials import AzureKeyCredential
endpoint = "https://models.inference.ai.azure.com"
model_name = "Mistral-large"
token = os.environ["GITHUB_TOKEN"]
client = ChatCompletionsClient(
endpoint=endpoint,
credential=AzureKeyCredential(token),
)
response = client.complete(
messages=[
SystemMessage(content="You are a helpful coding assistant."),
UserMessage(content="Can you write a Python function to the fizz buzz test?"),
],
temperature=1.0,
top_p=1.0,
max_tokens=1000,
model=model_name
)
print(response.choices[0].message.content)เมื่อเทียบกับสองโมเดลที่กล่าวถึงในบทเรียนนี้ Mistral NeMo เป็นโมเดลเดียวที่ฟรีและมีใบอนุญาต Apache2
โมเดลนี้ถือเป็นการอัปเกรดจาก LLM โอเพนซอร์สรุ่นก่อนของ Mistral คือ Mistral 7B
คุณสมบัติอื่น ๆ ของโมเดล NeMo ได้แก่
-
การตัดคำที่มีประสิทธิภาพมากขึ้น: โมเดลนี้ใช้ Tekken tokenizer แทน tiktoken ที่ใช้กันทั่วไป ซึ่งช่วยให้ทำงานได้ดีขึ้นกับหลายภาษาและโค้ด
-
การปรับแต่งเพิ่มเติม (Finetuning): โมเดลฐานพร้อมให้ปรับแต่งเพิ่มเติมได้ ซึ่งเพิ่มความยืดหยุ่นสำหรับกรณีการใช้งานที่ต้องการการปรับแต่ง
-
Native Function Calling - เช่นเดียวกับ Mistral Large โมเดลนี้ได้รับการฝึกฝนให้รองรับการเรียกฟังก์ชันโดยตรง ซึ่งทำให้เป็นหนึ่งในโมเดลโอเพนซอร์สรุ่นแรกที่มีฟีเจอร์นี้
ในตัวอย่างนี้ เราจะดูว่า Mistral NeMo จัดการการตัดคำอย่างไรเมื่อเทียบกับ Mistral Large
ตัวอย่างทั้งสองใช้คำสั่งเดียวกัน แต่คุณจะเห็นว่า NeMo คืน token น้อยกว่า Mistral Large
pip install mistral-common# Import needed packages:
from mistral_common.protocol.instruct.messages import (
UserMessage,
)
from mistral_common.protocol.instruct.request import ChatCompletionRequest
from mistral_common.protocol.instruct.tool_calls import (
Function,
Tool,
)
from mistral_common.tokens.tokenizers.mistral import MistralTokenizer
# Load Mistral tokenizer
model_name = "open-mistral-nemo "
tokenizer = MistralTokenizer.from_model(model_name)
# Tokenize a list of messages
tokenized = tokenizer.encode_chat_completion(
ChatCompletionRequest(
tools=[
Tool(
function=Function(
name="get_current_weather",
description="Get the current weather",
parameters={
"type": "object",
"properties": {
"location": {
"type": "string",
"description": "The city and state, e.g. San Francisco, CA",
},
"format": {
"type": "string",
"enum": ["celsius", "fahrenheit"],
"description": "The temperature unit to use. Infer this from the users location.",
},
},
"required": ["location", "format"],
},
)
)
],
messages=[
UserMessage(content="What's the weather like today in Paris"),
],
model=model_name,
)
)
tokens, text = tokenized.tokens, tokenized.text
# Count the number of tokens
print(len(tokens))# Import needed packages:
from mistral_common.protocol.instruct.messages import (
UserMessage,
)
from mistral_common.protocol.instruct.request import ChatCompletionRequest
from mistral_common.protocol.instruct.tool_calls import (
Function,
Tool,
)
from mistral_common.tokens.tokenizers.mistral import MistralTokenizer
# Load Mistral tokenizer
model_name = "mistral-large-latest"
tokenizer = MistralTokenizer.from_model(model_name)
# Tokenize a list of messages
tokenized = tokenizer.encode_chat_completion(
ChatCompletionRequest(
tools=[
Tool(
function=Function(
name="get_current_weather",
description="Get the current weather",
parameters={
"type": "object",
"properties": {
"location": {
"type": "string",
"description": "The city and state, e.g. San Francisco, CA",
},
"format": {
"type": "string",
"enum": ["celsius", "fahrenheit"],
"description": "The temperature unit to use. Infer this from the users location.",
},
},
"required": ["location", "format"],
},
)
)
],
messages=[
UserMessage(content="What's the weather like today in Paris"),
],
model=model_name,
)
)
tokens, text = tokenized.tokens, tokenized.text
# Count the number of tokens
print(len(tokens))หลังจากจบบทเรียนนี้แล้ว ลองดู Generative AI Learning collection เพื่อพัฒนาความรู้ด้าน Generative AI ของคุณให้ก้าวหน้าต่อไป!
ข้อจำกัดความรับผิดชอบ:
เอกสารนี้ได้รับการแปลโดยใช้บริการแปลภาษาอัตโนมัติ Co-op Translator แม้เราจะพยายามให้ความถูกต้องสูงสุด แต่โปรดทราบว่าการแปลอัตโนมัติอาจมีข้อผิดพลาดหรือความไม่ถูกต้อง เอกสารต้นฉบับในภาษาต้นทางถือเป็นแหล่งข้อมูลที่เชื่อถือได้ สำหรับข้อมูลที่สำคัญ ขอแนะนำให้ใช้บริการแปลโดยผู้เชี่ยวชาญมนุษย์ เราไม่รับผิดชอบต่อความเข้าใจผิดหรือการตีความผิดใด ๆ ที่เกิดจากการใช้การแปลนี้